The project was a traffic management system that worked on the principle of Real-Time Bidding from various advertising providers. In addition to the fact that the system had to have a high speed of response to a request - about 100 milliseconds, these 100 milliseconds should also include time to decide whether or not to bet on showing ads at this particular moment and if so then at what price. This decision was made based on the analysis of previous statistics on the effectiveness of advertising platforms according to a set of parameters, for example, the user's country, operating system, time of day, and so on. The created system had to withstand about 10,000 requests per second. Such a huge number of requests generated an ever-increasing array of data, which became almost impossible to analyze in real-time to make the right decision on the rate since statistical indicators often had to be recalculated preferably in real-time. Scaling up standard relational databases would ultimately entail large financial costs for maintaining and managing the infrastructure, and it would cost about $ 10,000 per month. As a result, after a series of tests, it was proposed to use the column-oriented database ClickHouse. Since this solution was open-source and at the same time used the usual SQL query language, it also had the necessary stability and speed of implementation, its implementation allowed to cut hosting costs up to $ 500 per month, which allowed saving $ 9,500 every month, meeting basic requirements to the system in terms of speed and stability of the response based on constantly incoming new data.